<![CDATA[The understanding of murottal Al-Quran among the Indonesian population remains relatively limited. One contributing factor is the difficulty in distinguishing between different murottal rhythms, which requires specialized expertise. Additionally, traditional murottal learning methods necessitate direct interaction with expert teachers, which is not always accessible to everyone. These challenges highlight the importance of developing technology to assist in identifying murottal rhythms. This study developed a murottal rhythm classification model using Convolutional Neural Networks (CNN) with transfer learning, employing two popular architectures: VGG16 and ResNet50. Audio data were processed using Short-Time Fourier Transform (STFT) and Mel-Frequency Cepstral Coefficients (MFCC) feature extraction for analysis.The results showed that the ResNet50 architecture with MFCC-extracted data achieved the best performance, with a training accuracy of 92%, validation accuracy of 85%, and testing accuracy of 86%. Additionally, the model achieved precision, recall, and F1-score values of 0.87 and 0.86, indicating strong generalization capabilities. Conversely, the VGG16 architecture with STFT and MFCC-extracted data demonstrated lower accuracy compared to ResNet50. The findings are expected to provide an innovative solution for developing a self-learning system based on technology to facilitate understanding of murottal rhythms in the Al-Quran.]]>
Copyrights © 2025
